RESUMEN
Artificial intelligence can potentially provide a substantial role in streamlining chest computed tomography (CT) diagnosis of COVID-19 patients. However, several critical hurdles have impeded the development of robust AI model, which include deficiency, isolation, and heterogeneity of CT data generated from diverse institutions. These bring about lack of generalization of AI model and therefore prevent it from applications in clinical practices. To overcome this, we proposed a federated learning-based Unified CT-COVID AI Diagnostic Initiative (UCADI, http://www.ai-ct-covid.team/), a decentralized architecture where the AI model is distributed to and executed at each host institution with the data sources or client ends for training and inferencing without sharing individual patient data. Specifically, we firstly developed an initial AI CT model based on data collected from three Tongji hospitals in Wuhan. After model evaluation, we found that the initial model can identify COVID from Tongji CT test data at near radiologist-level (97.5% sensitivity) but performed worse when it was tested on COVID cases from Wuhan Union Hospital (72% sensitivity), indicating a lack of model generalization. Next, we used the publicly available UCADI framework to build a federated model which integrated COVID CT cases from the Tongji hospitals and Wuhan Union hospital (WU) without transferring the WU data. The federated model not only performed similarly on Tongji test data but improved the detection sensitivity (98%) on WU test cases. The UCADI framework will allow participants worldwide to use and contribute to the model, to deliver a real-world, globally built and validated clinic CT-COVID AI tool. This effort directly supports the United Nations Sustainable Development Goals number 3, Good Health and Well-Being, and allows sharing and transferring of knowledge to fight this devastating disease around the world.
RESUMEN
BackgroundTo date, large amounts of epidemiological and case study data have been available for the Coronavirus Disease 2019 (COVID-19), which suggested that the mortality was related to not just respiratory complications. Here, we specifically analyzed kidney functions in COVID-19 patients and their relations to mortality. MethodIn this multi-centered, retrospective, observational study, we included 193 adult patients with laboratory-confirmed COVID-19 from 2 hospitals in Wuhan, 1 hospital in Huangshi (Hubei province, 83 km from Wuhan) and 1 hospital in Chongqing (754 km from Wuhan). Demographic data, symptoms, laboratory values, comorbidities, treatments, and clinical outcomes were all collected, including data regarding to kidney functions. Data were compared among three groups: non-severe COVID-19 patients (128), severe COVID-19 patients (65) and a control group of other pneumonia (28). For the data from computed tomographic (CT) scans, we also included a control group of healthy subjects (110 cases, without abnormalities in the lung and without kidney diseases). The primary outcome was a common presence of kidney dysfunctions in COVID-19 patients and the occurrence of acute kidney injury (AKI) in a fraction of COVID-19 patients. Secondary outcomes included a survival analysis of COVID-19 patients in conditions of AKI or comorbid chronic illnesses. FindingsWe included 193 COVID-19 patients (128 non-severe, 65 severe (including 32 non-survivors), between January 6th and February 21th,2020; the final date of follow-up was March 4th, 2020) and 28 patients of other pneumonia (15 of viral pneumonia, 13 of mycoplasma pneumonia) before the COVID-19 outbreak. On hospitaladmission, a remarkable fraction of patients had signs of kidney dysfunctions, including 59% with proteinuria, 44% with hematuria, 14% with increased levels of blood urea nitrogen, and 10% with increased levels of serum creatinine, although mild but worse than that in cases with other pneumonia. While these kidney dysfunctions might not be readily diagnosed as AKI at admission, over the progress during hospitalization they could be gradually worsened and diagnosed as AKI. A univariate Cox regression analysis showed that proteinuria, hematuria, and elevated levels of blood urea nitrogen, serum creatinine, uric acid as well as D-dimer were significantly associated with the death of COVID-19 patients respectively. Importantly, the Cox regression analysis also suggested that COVID-19 patients that developed AKI had a [~]5.3-times mortality risk of those without AKI, much higher than that of comorbid chronic illnesses ([~]1.5 times risk of those without comorbid chronic illnesses). InterpretationTo prevent fatality in such conditions, we suggested a high degree of caution in monitoring the kidney functions of severe COVID-19 patients regardless of the past disease history. In addition, upon day-by-day monitoring, clinicians should consider any potential interventions to protect kidney functions at the early stage of the disease and renal replacement therapies in severely ill patients, particularly for those with strong inflammatory reactions or a cytokine storm. FundingNone.
RESUMEN
Objective@#To investigate the value of combined chemical exchange saturation transfer (CEST) and conventional magnetizationtransfer imaging (MT) in detecting metabolic and structural changes of renal fibrosis in rats with unilateral ureteral obstruction(UUO) at 3T MRI. @*Materials and Methods@#Thirty-five Sprague-Dawley rats underwent UUO surgery (n = 25) or sham surgery (n = 10). Theobstructed and contralateral kidneys were evaluated on days 1, 3, 5, and 7 after surgery. After CEST and MT examinations,18F-labeled fluoro-2-deoxyglucose positron emission tomography was performed to quantify glucose metabolism. Fibrosis wasmeasured by histology and western blots. Correlations were compared between asymmetrical magnetization transfer ratio at1.2 ppm (MTRasym(1.2ppm)) derived from CEST and maximum standard uptake value (SUVmax) and between magnetization transferratio (MTR) derived from MT and alpha-smooth muscle actin (α-SMA). @*Results@#On days 3 and 7, MTRasym(1.2ppm) and MTR of UUO renal cortex and medulla were significantly different from those ofcontralateral kidneys (p < 0.05). On day 7, MTRasym(1.2ppm) and MTR of UUO renal cortex and medulla were significantly differentfrom those of sham-operated kidneys (p < 0.05). The MTRasym(1.2ppm) of UUO renal medulla was fairly negatively correlated withSUVmax (r = -0.350, p = 0.021), whereas MTR of UUO renal medulla was strongly negatively correlated with α-SMA (r = -0.744, p <0.001). @*Conclusion@#CEST and MT could provide metabolic and structural information for comprehensive assessment of renal fibrosisin UUO rats in 3T MRI and may aid in clinical monitoring of renal fibrosis in patients with chronic kidney disease.
RESUMEN
Objective To investigate the expression and distribution of Cdc42-interacting protein 4 (CIP4) in renal fibrotic tissue,and the interaction between CIP4 and β-catenin in transforming growth factor β1 (TGF-β1)-induced epithelial-mesenchymal transition (EMT) model of HK-2 cell line.Methods In vivo,the model of renal fibrosis was induced by 5/6 subtotal nephrectomy in rat.Masson staining was used to evaluate the level of renal tissue fibrosis.The expression and distribution of CIP4 was detected by immunohistochemistry.In vitro,the EMT model of HK-2 cell line was induced by TGF-β1 (10 μg/L) for 72 h.Western blotting was used to observe the expression of E-cadherin,β-catenin,CIP4 and α-SMA.Colocalization and interaction of CIP4 and β-catenin were detected by immunofluorescence and immunoprecipitation respectively.Results Compared to sham group,CIP4 expression was increased in group of 5/6 subtotal nephrectomy,CIP4 was mainly distributed in basolateral side of renal tubular epithelia.In vitro,expressions of α-SMA and CIP4 were increased in HK-2 cells stimulated by TGF-β1 for 72 h (2.5and 1.8 folds,respectively) (all P<0.05),expression of E-cadherin was decreased(P<0.05).Partial colocalization between CIP4 and β-catenin was detected by immunofluorescence.In control group,CIP4 and β-catenin partially colocalized at the cell membrane.Mter the stimulation of TGF-β1,translocation to nucleus of CIP4 and β-catenin were increased,and partially colocalized in nucleus.The interaction between CIP4 and β-catenin was observed by immunoprecipitation in both control and TGF-β1 stimulated groups.Conclusions Expression of CIP4 in renal fibrotic tissue is increased,which is mainly distributed in basolateral side of renal tubular epithelia.CIP4 and βcatenin partially colocalize and interact with each other.CIP4 may play a role in EMT process through the interaction with β-catenin.
RESUMEN
Objective To observe the effect of CIP4(Cdc42 interacting protein 4)on human renal tubular epithelial to mesenchymal transition(EMT)induced by transforming growth factor β1(TGF-β1)and to study the associated mechanism. Methods Human proximal tubular epithelial cells (HK-2 cell line) were cultured with TGF-β1 (10μg/L) for 72 hours. The protein expressions of E-cadherin and α-SMA were measured by Western blotting. One set of siRNA oligos specific for CIP4 and CIP4 construction of the entire coding sequence were designed based on the full CIP4 sequence in GenBank. Then HK-2 cells were transfected with CIP4-siRNA or pcDNA3.1-hCIP4 via lipofactamine 2000. The protein expressions of CIP4, E-cadherin and α-SMA were evaluated respectively in control cells, TGF-β1 treated cells, siRNA transfected cells, pcDNA3.1-hCIP4-transfected cells by Western blotting. The distribution of E-cadherin and α-SMA was observed by confocal microscope. After TGF-β1-treated HK-2 cells were interferenced with specific inhibitor of PI3K-Akt (wortmannin) 1μmol/L for 48 hours, Western blotting was used to detect the CIP4 protein in control cells and interferenced cells. Results With TGF-β1 stimulation, the expression of E-cadherin protein was decreased markedly (P<0.05), and in contract, the expression of α-SMA were increased notably (P<0.05), which revealed that TGF-β1 could induce EMT. After transfected with CIP4-siRNA, the protein expression of E-cadherin was increased (P<0.05), and the protein expression of α-SMA was decreased (P<0.05). The EMT induced by TGF-β1 was effectively reversed. After transfected with pcDNA3.1-hCIP4, the expression of E-cadherin protein was down-regnlated (P<0.05), and the expression of α-SMA protein was up-regulated compared with control group (P<0.05), leading to EMT. After HK-2 cells were interferenced with wortmannin for 48 hours, the expression of CIP4 was decreased (P<0.05). Conclusion TGF-β1 upregulates the expression of CIP4 via PI3K-Akt pathway, and CIP4 may participate in EMT induced by TGF-β1.
RESUMEN
Objective To investigate the expression of Erbin in renal interstitial fibrosis (RIF) and the effect of over-expression of Erbin on transforming growth factor β1 (TGF-(β1)-induced epithelial-mesenchymal transition (EMT) in NRK52E cells. Methods In vivo, the model of renal fibrosis was induced by 5/6 subtotal nephrectomy in rat. Scr and BUN was detected and Masson staining was used to evaluate the level of renal tissue fibrosis. The location and expression of Erbin in renal tissue were detected by immunohistochemistry and Western blotting. In vitro, after NRK52E cells were treated by TGF-β1 (10 μg/L) for 72 h, immunofluorescence and Western blotting were used to obverse the expression and distribution of E-cadherin and α-SMA. The expression of Erbin mRNA and protein were detected by RT-PCR and Western blotting respectively. NRK52E cells were transiently transfected with Prk5-myc-Erbin plasmid via lipofectamine 2000, then the expressions of Erbin, E-cadherin and α-SMA were detected by Western blotting. Results (l)Compared to sham group with Scr (33.96±7.28) μmol/L and BUN (8.11±2.55) mmol/L, rats in 5/6 nephrectomy model with Scr (140.52±61.11) μmol/L and BUN (34.23±7.66) mmol/L revealed renal dysfunction. Masson staining indicated kidney interstitial fibrosis, and the expression of Erbin was significantly increased in renal tissue(2.9 folds), especially in tubular epithelia. (2)In vitro, the expressions of Erbin and α-SMA were markedly increased (2.3 folds and 2.1 folds, P<0.05, respectively) and the expression of E-cadherin was dramatically decreased in NRK52E cells stimulated by TGF-β1, which were consistent with immunofluorescence results. TGF-β1-induced E-cadherin suppression and a-SMA induction could be efficiently blocked by over-expression of Erbin (all P <0.05). Conclusions Erbin is up-regulated in renal interstitial fibrosis, and over-expression of Erbin can partly inhibit renal EMT induced by TGF-β1, which indicates Erbin playing an protective role in renal fibrosis.
RESUMEN
Objective To observe the expression and localization of CIP4 (Cdc42 interacting protein-4) in the renal fibrosis and the effect of CIP4 on the expression of E-cadherin,vimentin and β-catenin tyrosine phosphorylation. Methods In vitro, the human tubular epithelial cells (HK-2 cell line) were cultured with 10 μg / L TGF-β1 for 72 h. The protein expressions of CIP4, E-cadherin, vimentin and β-catenin tyrosine phosphorylation were measured by Western blotting; the expression of CIP4 mRNA was detected by RT-PCR. The intracellular distribution of CIP4 was observe by confocal microscope. In vivo, Masson staining was used to evaluate the level of renal fibrosis; the expression and distribution of CIP4 in renal tissue were detected by immunohistochemistry. HK-2 cells were transfected with pcDNA3. 1-CIP via lipofectamine 2000. The expressions of E-cadherin, vimentin and β-catenin tyrosine phosphorylation level in the transfected cells were detected by Western blotting. Results The expressions of CIP4 mRNA and protein were up-regulated in renal tubular EMT cells. Most of CIP4 protein localized in cell membrane, and some was in cytoplasm. After stimulation by TGF-β1, the expression of CIP4 protein both in cytoplasm and nucleus was greatly increased (P <0.05),especially in cytoplasm. In vivo, CIP4 was expressed in renal tubular epithelia, but little expressed in glomeruli. In renal from 5/6 nephrectomized rats, CIP4 expression was significantly increased. In the CIP4 transfectants, the expression of CIP4, vimentin and β-catenin tyrosine phosphorylation level were up-regulated (P <0.05), but E-cadherin expression was suppressed (P <0.05).Conclusion The overexpression of CIP4 is likely to take part in the epithelial-to-mesenchymal transition process, thereby promoting the renal fibrosis.